LINK AND BEAM Connection

[keler]

I'm from Poland and I try to deal with Ansys. My name is Piotr. I'd like to
ask a question-I suppose It is rather simple for You to answer.
So, let me describe the problem: I need do connect LINK Element with
BEAM element in 3D analysis and as You maybe spotted DOF's are
inconsistent (UX,UY,UZ as for LINK and UX,UY,UZ,ROZX,ROTY,ROTZ as
for BEAM). So ma question is how to cope with that? Is the only
solution to use constrain equations? (Let's suppose I need swivel
joint between LINK and Beam).
Thanks in advance!!!

 

[mastroit2002]

The problem you mentioned could be solved using a foot like end at Your beam element end.

Since only the displacements, not rotations, are the unknows on the general description of "normal" 3D element, one has to simulate generalized ratations via a differential displacements.

The beam end should be connected rigidly and orthogonaly to not less than 3 nodes non lieing on a line.

The concern of local stress would be less important.

Hope this will be of use.

mastro

 

[keler]

OK not bad solution, but let's go further. How about Link and Shell in 3D (2D whatever)????? Suppose I must support an engine and I need to consider swivel joints as connections??? Are there rules to connect elements witch DOF's are inconsistent???
Best Regards.
Peter "Keler" from Poland.

 

[mastroit2002]

>How about Link and Shell in 3D (2D whatever)?????

I'm not sure to get correctly what You mean: let's assume You do want to connect a 2D element shell like element with a 3D Lagrangian element (h-element not p-like element [Lagrangian vs Hermitian one] so only displacements are the unknowns).

>Suppose I must support an engine and I need to consider >swivel joints as connections???

So You claim that there could be a relative rotation at knots: so there are several possibilities.

The knots of connection lie on a line so relative rotation is possible with line as normal of rotation. Otherwise the knots doesn't lie on more than one simple straight line -> the connection will use the shell own stiffness to effort the joint and so on..

More generally one should give the insight to a FEM program to perform a suitable calculation: in this sense a mixed element type problem is sometime difficult to analyse.

Mastro

P.S. I tried with MSC software a case of junction between tetra and beam elements and differently from my opinion the problem even if very ill conditioned was solved: one should be even more careful.